Piketty's Three Big Mistakes | Hacker News Reader

Piketty's Three Big Mistakes

bloombergview.com

66 pointstptacek67 comments

You can't say the rich have lots of land, then complain about not taking depreciation into account. Land doesn't depreciate. This guy is just searching for reasons to dismiss Picketty.

Edit: If you didn't read the article: "Rognlie points out that almost all of the increase in the value of capital over Piketty’s timeline comes from land"

http://www.accountingcoach.com/blog/why-isnt-land-depreciate...

Yea, a truly great economics troll would reach for something like the money multiplier, law of diminishing marginal returns, or loanable funds to make such an argument.

Or at least touch on the rise of the modern un-backed central bank post Bretton Woods. Rich people can more easily shield their wealth from inflation than middle class or poor people can shield their wages from inflation, leading to a perpetually increasing gap between the two so long as central banks continue to use QE and inflationary schemes to fake economic growth and dilute debt. Average workers earning 1% annual wage increases don't tend to do very well when you debase their currency at over twice that rate. A good economic troll would point that out.

QE wasn't/isn't inflationary (and I believe the empirics bear this out), it was a dollar for dollar swap of treasuries for reserves.

If anything it is/was the opposite, removing interest income from the private sector and putting it on the FED's books (which get remitted back to the treasury).

(ooh, you're goooood)

There is a long-standing distinction within classical political economy of differentiating between land, labor, and capital holding classes. Adam Smith, David Ricardo, and Karl Marx among many others make this distinction.

Capital holders face depreciation. Land holders do not (at least, not to the same degree). Piketty looks at the asset mix of the rich over time, arguing that the rich have predominantly shifted away from land holding toward capital holding. Pre-WWI most wealth was drawn from land holdings, whereas Post-WWII most wealth is drawn from capital holdings and super incomes.

The argument about the reconcentration of wealth from the 70s and 80s through the present ought to be affected if depreciation of capital were much higher than Piketty allowed for in his model—especially since the rich mostly hold capital now.

If the wealth increase over time has been coming from land and not capital (which faces depreciation, while land does not), that's an interesting and important fact, because it affects the policy recommendations that flow from Piketty's argument. Piketty's conclusions call for a global progressive wealth tax and a global progressive income tax. What the article points out is that if most of the wealth increase is coming from land, a different type of tax might be the best way to target the concentration of wealth—namely the Georgist land-value tax instead of broader global wealth and income taxes.

It doesn't appear that the MIT student is just casting about to find reasons to dismiss Piketty, rather these may be significant insights into the data that affect the overall conclusions that we might draw from it. Just like Piketty, Rognlie should be judged on the quality of his data.

Discussing his solutions is very different from discussing his evidence.

His evidence is sound but that does not mean his conclusions are or need to be. IMO you could throw away the entire last part of the book and still have one of the most important books written about capitalism.

Now all we need is someone who write one about technology which is the elephant in the room and why I think Piketty is wrong about his proposals but thats another story.

I agree that his evidence is sound. But like any other constructed data set, Piketty makes choices about measurement indicators and categories that are open to challenge. The way that he composes his measures of wealth and capital could be more nuanced, if those nuances make a significant difference in the results produced by the models built on top of them. This MIT student seems to be doing just that—arguing that the composition of land and capital in measures of wealth need to be treated with more nuance, and that measures of capital need to take account of depreciation in a different way than Piketty does.

I make similar arguments in my dissertation about one of the longest-running, most widely used data sets in my field. Sometimes the categorization of data is misleading in certain ways that elides phenomena that are actually very significant.

Piketty's construction of the data set is fantastic work. But it doesn't mean that all the choices he made in constructing and categorizing the data are the best choices that could have been made.

You are probably right about that but I don't think better choices would change the result, but thats just my personal opinion.

"What the article points out is that if most of the wealth increase is coming from land, a different type of tax might be the best way to target the concentration of wealth—namely the Georgist land-value tax instead of broader global wealth and income taxes."

I hate to be Mr. Unintended Consequences, blah blah, but ...

Can we stop and imagine for a bit what the world would look like if land owners were forced to build and develop to match rising land values just to maintain their ownership ?

Right now we sometimes see long time property owners forced to sell - they cannot afford to live there anymore as property taxes rise, based on the value. Imagine if the buildings themselves (or the farms or the gardens or the parks) could not afford to be there ?

Hmm... Sure you can?

Non-land domestic wealth is still a significant share and his arguments for accelerated depreciation ring true to me.

Read the article.

"Rognlie points out that almost all of the increase in the value of capital over Piketty’s timeline comes from land"

Look at the data.

Non-land wealth is still huge, and he's arguing the nature of this wealth brings higher depreciation than it once did.

A direct quote from Rognlie's article:

"Land does not depreciate; all its income is net."

This is the worst kind of internet debate: a group of people who haven't read Piketty's book arguing with a group of people who haven't read an article which claims to find problems with it.

I think many economists want to discredit him because he points out how many of them are paid by special interests.

BTW, I liked Capital in the 21st Century - a good dead.

Fancy searching for reasons to dismiss someone's argument or thesis! That's not playing fair at all.

Pikketys contribution is not his conclusion but the data that backs up the conclusion.

Most people already knew what Pikkety concluded, they just didn't have evidence for it. Now we do and until someone shows better evidence it's hard to claim Piketty is making mistakes without most probably being politically motivated to do so.

His solution on the other hand was an unnecessary but understandable addition to the book.

And I say this as a European style liberal.

Please correct the typo note for not; because this is most correct thing that can be said about Pikkety's work.

His conclusions are arguable, and can even be dismissed. But the data he collected and organized can and should be taken seriously and reworked. Yes, revised, treated with different assumptions in mind. But dismiss his whole work because he might be wrong on his final politic suggestions is a disservice to the science (no matter how small amount of science there is in Economics).

If you want to make Economics learn some things from hard sciences, you should take Piketty's data seriously.

Corrected and thanks for the catch!

> But dismiss his whole work because he might be wrong on his final politic suggestions is a disservice to the science (no matter how small amount of science there is in Economics).

Not in the least because he himself appears to be very open to criticism, as a proper scientist should be.

Unless the data itself is suspect, which it is.

http://www.ft.com/cms/s/0/c9ce1a54-e281-11e3-89fd-00144feabd...

http://www.nytimes.com/2014/05/24/upshot/did-piketty-get-his...

http://www.wsj.com/articles/alan-reynolds-why-pikettys-wealt...

It's interesting to see how many people so dearly want his assertions to be true.

Not at least you :)

Those claims have been discussed here before and they don't change anything substantial.

>it's hard to claim Piketty making mistakes without most probably being politically motivated to do so.

Precisely. The media and punditry love to emphasize that economic models didn't see this or that crash, or can't explain this or that phenomena. Yet this guy writes an incredibly well researched, deep book, and the same people pile on because his conclusions don't model reality to a tee. Economics doesn't work that way. All researchers can do is attempt to get incrementally closer to "the truth".

We're going through some interesting times of awareness of wealth inequality globally, so both the "socialist" side and "capitalist" side have exaggerated his conclusions, and thus the guy can't stay out of the news.

Piketty's big contribution is the creation of time series of top wealth distribution going back to the ~1850. In the linked paper, all the time series start at 1960 at the earliest, thus dismissing most of the new empirical data.

So whatever technical merit the critique has, it doesn't seem to concern itself with the big picture results since it neglects the empirical work.

Piketty's thesis is that the concentration of wealth in the permanent upper class of the 1800's is the historical law, and that the mid-20th Century prosperity of the middle class is a historical anomaly that was created by unprecedented destruction of real property by the two world wars, and further that now that the world wars are receding into history, we are returning to the 19th Century status quo.

If post-1960 data does not show a return to the status quo, then that undermines the core of Piketty's claims. Nobody is claiming that the 19th Century had a strong middle class.

Economic and political conditions of the mid 19th century led directly to the development of ideologies that we fought wars over in the first half of the 20th century.

I've always wondered whether humanity would enter a new cycle of competing ideologies. Maybe we are...

None of these arguments seems very strong. I don't believe that depreciation costs are significant compared to the rate of growth in income among wealthy individuals. No numbers are presented to support that hypothesis and intuitively it seems unlikely. I don't see how the third point about landlords cuts against Piketty's premise at all; on the contrary, it supports it.

Only the second point has some merit. It is possible that we're not measuring real wealth growth, but merely the froth in an economic bubble. But again, I'm not sure that historically that froth accounts for the disparity. Despite the correction in 2008, the wealthy are still handily beating GDP growth over the past 20 years.

you don't have to go by the blog post, but rather you can read Matthew Rognile's paper, which goes into this in detail

http://www.brookings.edu/~/media/Projects/BPEA/Spring-2015/2...

I thought the land case was implicitly part of Piketty's critique already?

The equitable way to stop the rich getting disproportionately richer is to tax other ways of gaining wealth - in particular investment income and changes in property value - the same way we tax labour income.

The mainstream Standard Model of economics generally replies to calls like yours with:

1. Investment income has already been taxed once labour income, it isn't "fair" to tax it "again".

2. Property value taxes drive out elderly on fixed income.

Soros' promotion of economic reflexivity broadly addresses these responses. As a practical matter however, capital behind investment income (including property interests) has generally captured sufficiently large sections of political, legislative and regulatory infrastructures that you could theoretically muster enough votes to get taxes ostensibly "raised", yet still get defeated in detail when it gets down to brass tacks implementation in those areas.

The way it works is if the political races cannot be tilted in favor of sufficient numbers of legislators to dilute/destroy the implementation of the vote through successive election cycles, then pressure through lobbying will attempt to sway legislative processes (from standard lobbying tactics to clamoring for parliamentary manoeuvres, etc.), and if the changes get through that, then regulatory capture will work on diluting/nullifying enforcement.

It's not hopeless; go into it with eyes wide open.

I don't see how any of these points directly threaten piketty's conclusion, that "r > g is a force for divergence," at least as they are presented here.

The first counterargument is nothing new. It's in the Solow model for Pete's sake! It only threatens the returns to capital as a theoretical conclusion in the distant, long-run future (going off the Solow model). Piketty has made an empirical argument off historical data.

The second paragraph saying most of r has come from capital gains in the financial market just looks silly before the one saying most of capital gains has been from an appreciation in land value. I don't know if that is how the original argument goes.

r > g is definitely a force for divergence. my intuition was that it is empirically unrealistic to expect r > g and Rognile's paper seems to provide quantitative evidence for this intuition

I explain many of the same things, but I don't think Piketty is 'wrong'. It is our responsibility to address the current preference for 'on paper' capital and keep it from getting out of control.

https://vimeo.com/user17783424/review/122784294/439ab072b5 <-fixed

Your vimeo link is broken for me.

Thanks for the catch...I don't know why they do this to pro users.

If these effects are strong enough to undermine Piketty's hypothesis then it is incumbent upon Rognlie to run the numbers and show that they do otherwise it's just more fud.

Here's a link to the paper if you're interested:

http://www.mit.edu/~mrognlie/piketty_diminishing_returns.pdf

Most people just read Piketty to page 26 anyway. So anyone claiming they read it, probably not.

The Summer's Most Unread Book Is… http://www.wsj.com/articles/the-summers-most-unread-book-is-...

""Capital in the Twenty-First Century" by Thomas Piketty: 2.4% Yes, it came out just three months ago. But the contest isn't even close. Mr. Piketty's book is almost 700 pages long, and the last of the top five popular highlights appears on page 26. Stephen Hawking is off the hook; from now on, this measure should be known as the Piketty Index."

It seems to be some kind of new sport for economists and economy journalists: Finding some errors in Pikettys book and writing about it as they had found the stone of wisdom.

I'm not sold that the overhead of being a capitalist has risen as drastically as is made out. Obsolete isn't the same as broken, no matter how low the resale value.

You have to appreciate the impact that Piketty has made on economics. It's one year later and columnists and academics are writing about the blog post of students that have minor quibbles with his data or conclusions.

The original paper that Rognlie wrote:

http://www.mit.edu/~mrognlie/piketty_diminishing_returns.pdf

This is the story of how a grad student's blog comment at Marginal Revolution appears to have blown up into one of the most influential critiques of Piketty to date. The WaPo story linked here has more more detail as well.

Piketty's argument that capital will grow at a rate significantly faster than GDP should seem weak to pretty much anyone who has ever tried to invest reasonable size capital at a decent return. Think about it. How can returns outpace gdp growth for a very long time ( obviously it can happen over a short time horizon ) ? It would require some kind of magic bean that as an investor, I would love to get a hold of

the person downvoting, be interested to hear an explanation of your reasoning :)

Why downvote instead of engage the argument? The S&P 500 return over the last 15 years has been pretty poor. VC returns have been poor. So, empirically, we don't seem to have this divergence in recent data.

(I did not downvote). But have you read Pikkety's book? It pretty thoroughly explains how capital grows more rapidly than GDP.

sure. My statement is that his mechanism for generating this growth appears to me to be unrealistic. New research appears to generate some confirmation for my view.

What research? Even over this anecdotally chosen timespan, the S&P 500 returned 3.98% / year on average. That completely squashes GDP growth over the same period.

no it doesn't. we've had 2 percent inflation over the same period.

> over this anecdotally chosen timespan, the S&P 500 returned 3.98% / year on average

> we've had 2 percent inflation over the same period

So, still, return on capital "completely squashes GDP growth over the same period" by twice as much, given your figure. No?

no, US real gdp growth and real S&P 500 returns have been basically in line in the 2000s. And real S&P 500 returns are way lower than world GDP growth over the same period.

Piketty assumes that technological innovations will raise returns on capital while not really raising GDP by the same amount. This is a very zero sum game view of the world, which I find strange. Do we really believe that rise of the robots will not increase our GDP ? self driving cars will not increase GDP ? better medical care will not increase GDP ? Maybe, but it doesn't seem realistic

Ah yes, sorry, I had a brain fart and confused growth and inflation...

As for Piketty's point on technological innovation, isn't it that it is one possible mechanism to escape the capital trap, but only if the rents on the resulting developments are low enough, and that historically they haven't been because of investor's desire for returns?

I think his point is that technological innovation will result in transfer of income from labor to investors that finance the innovation. It most certainly will, but he skirts the issue that the innovation itself could very well dramatically increase GDP !

Isn't it more that there will be a transfer in the share of income, rather than the absolute amount - i.e. not ruling out a rise in absolute GDP?

If you're adjusting for inflation, you should adjust both the GDP numbers and the stock indices.

of course. GDP growth numbers are always quoted in real terms

The magic bean is called 'congress'.

Reply on news.ycombinator.com